Clinical Utility and Feasibility of Smartphone-based EEG in Kenya: A Multicenter Observational Study

Anca Alina Arbune; Anton Mosquera Storgaard; Dylan Rice; Farrah J. Mateen; George Wambugu; Magnus Guldberg Pedersen; Melita Cacic Hribljan; Nomin Enkhtsetseg; Nshimiyimana Jules Fidele; Sandor Beniczky

arxiv: 2605.08157 · v3 · pith:SC5KCVJ4new · submitted 2026-05-04 · 📡 eess.SP · cs.CY

Clinical Utility and Feasibility of Smartphone-based EEG in Kenya: A Multicenter Observational Study

Nomin Enkhtsetseg , William Lehn-Schi{\o}ler , Anton Mosquera Storgaard , Magnus Guldberg Pedersen , Dylan Rice , George Wambugu , Nshimiyimana Jules Fidele , Melita Cacic Hribljan

show 4 more authors

Anca Alina Arbune Sidsel Armand Larsen Sandor Beniczky Farrah J. Mateen

This is my paper

Pith reviewed 2026-05-21 00:46 UTC · model grok-4.3

classification 📡 eess.SP cs.CY

keywords smartphone EEGfeasibility studyKenyaresource-limited settingsepilepsy diagnosispoint-of-care EEGnon-specialist operatorsLMICs

0 comments

The pith

Non-specialist operators in Kenya can acquire large-scale, interpretable EEG data using a smartphone-based system.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

This multicenter study tested a smartphone EEG device across 29 sites in Kenya to see if it could work in everyday clinical practice. Trained healthcare workers performed 3,036 recordings, and remote experts could interpret 96 percent of them. The system proved practical for common issues like seizures, with quick turnaround times for results. If this holds, it offers a way to bring brain wave testing to places without traditional EEG machines or specialists.

Core claim

The authors report that large-scale, point-of-care EEG acquisition by non-specialist operators in a resource-limited setting is feasible, as shown by 96% of 3,036 smartphone-based recordings being interpretable with a mean interpretation turnaround of 107 minutes.

What carries the argument

Smartphone-based 27-lead EEG system that allows trained healthcare workers to acquire standardized recordings for remote expert interpretation.

If this is right

High interpretability rates support use for diagnosing epilepsy and other neurological conditions.
Shorter recording times were linked to uninterpretable results, suggesting minimum duration guidelines.
Abnormal findings were common, especially epileptiform activity in children.
Mean turnaround time of 107 minutes enables timely clinical decisions.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

This approach could help close the diagnostic gap for epilepsy in sub-Saharan Africa by enabling more tests without building new hospitals.
Further studies comparing smartphone EEG directly to standard systems would strengthen confidence in the findings.
Scaling this could support broader neurological care in other low-resource countries.

Load-bearing premise

The smartphone-based recordings produce data of sufficient quality for reliable clinical interpretation by remote experts.

What would settle it

A head-to-head study where standard hospital EEG and smartphone EEG are performed on the same patients and the smartphone version misses clinically relevant abnormalities would challenge the claim.

read the original abstract

Purpose: Access to electroencephalography (EEG) remains limited across low- and middle-income countries (LMICs) due to cost, infrastructure requirements, and a shortage of trained staff. This study evaluated the feasibility and clinical utility of a smartphone-based EEG system in a real-world setting. Methods: We conducted a multicenter observational study (November 2023 to April 2026) across 29 clinical sites in Kenya. A smartphone-based 27-lead EEG system enabled trained healthcare workers to acquire standardized recordings with remote expert interpretation. Results: 3,036 EEG sessions were performed. Male patients constituted 57.8% of the cohort, with representation across pediatric and adult populations. The most common referral indication was seizures or convulsions (68.5%). Overall, 2,915 (96%) recordings were interpretable, while 121 (4%) were uninterpretable, primarily due to high electrode impedance and insufficient recording duration. Uninterpretable recordings were significantly shorter than interpretable recordings (mean 18.5 vs. 33.8 minutes; median 15.1 vs. 31.6 minutes; p < 0.0001). Mean turnaround time for interpretation was 107 minutes. Among interpretable recordings, 917 (30.2%) were abnormal, including 701 (76.4%) with epileptiform abnormalities, 215 (23.4%) with non-epileptiform findings, and 1 (0.1%) indeterminate finding. Epileptiform abnormalities were highest in children aged 4-9 years (33.1%) and less frequent in adults (14-21%). Non-epileptiform abnormalities were more common in patients aged 60+ years (19.2%) compared to younger age groups (3-9%). Conclusion: Large-scale, point-of-care EEG acquisition by non-specialist operators in a resource-limited setting is feasible. Expansion of smartphone-based EEG systems may improve equitable access to neurological diagnosis and care in LMICs.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

Smartphone EEG acquisition by non-specialists works at scale in Kenya with 96% interpretability, but lacks head-to-head validation against standard systems.

read the letter

The punchline is that this multicenter study shows smartphone EEG acquisition by non-specialists is feasible at scale in Kenya, with 96% of 3,036 recordings interpretable by remote experts. What is new is the large real-world deployment across 29 sites with actual clinical referrals from November 2023 to April 2026. The paper does well by giving specific performance numbers, including why recordings failed mainly due to high electrode impedance and insufficient duration, with a clear statistical comparison showing uninterpretable ones were much shorter (p < 0.0001). They also report a mean turnaround time for interpretation of 107 minutes and break down the 30% abnormal findings by type and age group. This provides useful benchmarks for similar efforts in other settings. The soft spots are the lack of a head-to-head quality comparison to conventional EEG and the observational nature without controls. These are not fatal for the feasibility claim, since interpretability serves as the proxy, but they do limit how far we can push the clinical utility angle without more data. The citation pattern seems appropriate for an applied study without over-relying on self-citations. This paper is for researchers and clinicians focused on improving diagnostic access in low- and middle-income countries. Readers interested in practical health technology implementation will get the most out of it. It has enough substance and transparency to deserve a serious referee. I recommend sending this to peer review.

Referee Report

0 major / 3 minor

Summary. The manuscript reports results from a multicenter observational study across 29 sites in Kenya (November 2023–April 2026) that deployed a smartphone-based 27-lead EEG system operated by trained non-specialist healthcare workers, with remote expert interpretation. Key findings from 3,036 recordings include a 96% interpretability rate, significantly shorter durations for the 4% uninterpretable cases (mean 18.5 vs. 33.8 min, p<0.0001), a mean interpretation turnaround of 107 minutes, and 30.2% abnormal recordings (76.4% of which showed epileptiform activity, highest in children aged 4–9 years).

Significance. If the results hold, the work provides large-scale empirical evidence that point-of-care EEG acquisition is feasible in resource-limited LMIC settings without specialist operators on site. Notable strengths include the multicenter design, high recording volume, and direct statistical support for technical factors affecting success; these elements strengthen the feasibility demonstration and its potential relevance for expanding neurological diagnostics.

minor comments (3)

[Results] Results section: sample sizes underlying the age-stratified abnormality percentages (e.g., 33.1% epileptiform in 4–9 years) are not reported, which limits assessment of precision and statistical power for those subgroup claims.
[Methods] Methods: the operational criteria used by remote experts to classify a recording as 'interpretable' versus 'uninterpretable' are described only at the level of failure reasons (impedance, duration); an explicit checklist or threshold would improve reproducibility.
[Discussion] The manuscript would benefit from a brief statement in the Discussion on how the observed abnormality rates compare with published conventional EEG cohorts in similar populations, even if only as context.

Simulated Author's Rebuttal

0 responses · 0 unresolved

We thank the referee for their positive and accurate summary of our multicenter observational study on smartphone-based EEG in Kenya, as well as the recommendation for minor revision. The significance assessment correctly highlights the value of the large-scale data, multicenter design, and statistical support for technical factors. No specific major comments were raised in the report.

Circularity Check

0 steps flagged

No significant circularity; purely empirical observational study

full rationale

The paper reports a multicenter observational study collecting 3,036 smartphone-based EEG recordings across Kenyan sites, with direct empirical metrics (96% interpretability rate, statistical comparisons of duration and impedance, turnaround times, and abnormality rates by age). No equations, derivations, predictive models, fitted parameters, or self-referential claims exist. All conclusions rest on observed data and remote expert interpretation as the quality proxy, without any reduction of results to inputs by construction or load-bearing self-citations. This is a standard non-circular empirical feasibility report.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 0 invented entities

The central claim rests primarily on empirical deployment data rather than new theoretical constructs; the main supporting elements are standard clinical assumptions about EEG interpretation.

axioms (1)

domain assumption Standard clinical EEG interpretation criteria remain valid when applied to recordings from the smartphone-based system.
Invoked when reporting abnormality rates and interpretability based on remote expert review without additional validation steps described.

pith-pipeline@v0.9.0 · 5982 in / 1271 out tokens · 65717 ms · 2026-05-21T00:46:53.837349+00:00 · methodology

Review history (2 revisions) →

discussion (0)

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

IndisputableMonolith/Foundation/RealityFromDistinction.lean reality_from_one_distinction unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

Large-scale, point-of-care EEG acquisition by non-specialist operators in a resource-limited setting is feasible.
IndisputableMonolith/Cost/FunctionalEquation.lean washburn_uniqueness_aczel unclear

?

unclear
Relation between the paper passage and the cited Recognition theorem.

96% recordings interpretable; failures due to impedance and duration (p<0.0001)

What do these tags mean?

matches: The paper's claim is directly supported by a theorem in the formal canon.
supports: The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends: The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses: The paper appears to rely on the theorem as machinery.
contradicts: The paper's claim conflicts with a theorem or certificate in the canon.
unclear: Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Forward citations

Cited by 3 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score.

Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders
cs.LG 2026-05 unverdicted novelty 6.0

Sparse autoencoders on EEG transformers identify three regimes of clinical concept encoding and reveal entanglements such as age-pathology confounding via a new steering selectivity metric.
Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders
cs.LG 2026-05 unverdicted novelty 6.0

TopK SAEs applied to EEG transformers extract clinical features, enable concept steering, and identify selectively steerable, entangled, and non-encoded regimes with a spectral decoder for physiological interpretation.
Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders
cs.LG 2026-05 unverdicted novelty 6.0

Sparse autoencoders on EEG transformers extract clinical features, identify three steering regimes, expose age-pathology entanglements and wrecking-ball failures, and map interventions to frequency spectra.